Module 14 · Score calls with quality management

Leer en español: Módulo 14 · Evalúa llamadas con calidad (QM).

Goal: build a scoring form, evaluate a recorded call, handle an agent's dispute and read the quality report. Time: ~8 minutes · You need: quality management enabled for your account, recordings of finished calls (Module 5), and a login with permission to evaluate (supervisor or admin seat, or account admin).

Quality management may be off for your account

Writing forms and evaluations only works when your account has quality management enabled. It is off by default and your Dialer Digital contact turns it on. While it is off, saving a form or an evaluation is refused with the message that QM is not enabled and nothing is saved. Evaluations that already exist stay readable.

1. Build a scoring form

Open Quality in the sidebar. The Quality forms panel lists your forms; choose New form.

1. Give it a Form name and a Pass threshold (%).
2. Add criteria with Add criterion. Each has a Stable key (letters, digits, _ or -; it never changes), a Prompt (the question the evaluator sees), an Answer type (Yes / no or Scale), a Weight and, for a scale, the Scale maximum.
3. Tick Allow N/A when a criterion may not apply to every call. Tick Critical on a yes/no criterion when a "no" must fail the whole evaluation (for example reading the legal disclosure).
4. Choose Save form. You will see "Saved as version 1."

Editing a form never rewrites the past: saving a change creates a new version, and evaluations already written stay tied to the version they were scored with. Archive form retires a form but keeps its evaluations readable. Only an account admin can create, edit or archive forms.

2. Evaluate a recording

Open Recordings and select a call; the Evaluate this recording panel appears below it. Only a recording that is available can be evaluated: for one that is still pending or was purged, saving is refused and nothing is written.

1. Pick the Form.
2. Answer every criterion (the panel asks you to answer them all before saving).
3. Write Feedback for the agent. The agent will read it.
4. Choose Save evaluation. The score is calculated by the server from the form version, and the panel shows "Evaluation saved: …%."

Things to know:

• You cannot evaluate your own call.
• A saved evaluation is never edited. To fix a mistake, evaluate again: Save correction adds a new evaluation and keeps the old one. The newer one is the one that counts.
• A critical criterion answered "no" fails the evaluation and its score is 0.

3. When the agent disputes

Agents see their evaluations in My evaluations on their screen: score, answers and your feedback. They do not see who evaluated. Inside the dispute window (7 days from the evaluation by default) the agent can choose a Reason (Wrong call, Criterion was misjudged or Other), explain what should be reviewed and choose Open dispute.

While a dispute is open, the evaluation keeps counting in the report, and you cannot correct it by re-evaluating. A supervisor or admin (never the evaluated agent) resolves it with one of two decisions:

• Upheld: the score stands.
• Adjusted: you provide new answers for every criterion and a new evaluation replaces the disputed one.

Resolution is done through the API in this version

The dashboard has no button to resolve a dispute yet. Resolve it through the public API: see Quality management, section "Resolve the dispute". The report shows the outcome as soon as it is saved.

4. Read the quality report

On Quality, in Quality reports, choose the Report dimension (By agent or By campaign), set From and To (at most 31 days) and choose Run report. For each agent or campaign and form you get: Evaluations, Average score, Pass rate, Critical failures and Open disputes.

• Only the current evaluations count: a correction replaces the one it corrects.
• Two forms are two different scales, so they appear as separate rows and are never averaged together.
• With no evaluations in the window the screen says "No evaluations in this window." and shows no average, never a false 0.

What you learned

• How to build a versioned scoring form with weights, N/A and critical criteria.
• How to evaluate a recording and correct it without losing history.
• How an agent's dispute works, and that it is resolved through the API in this version.
• How to read the report and why open disputes still count.

Developers: the full endpoint guide is Quality management.